Why Diffusion and Confusion prevent statistical analysis? [duplicate]

I've studied the concepts of Diffusion and Confusion of cryptographic algorithms, however, I don't understand why they help in preventing a statistical analysis. Are you able to explain to me?